Abstract
BACKGROUND: The associations between different definitions of prediabetes and diabetes with mortality and life expectancy remain unclear. Clarifying these associations is essential for informing hyperglycemia management policies. We aimed to compare the prevalence of prediabetes and newly diagnosed diabetes across different glycemic indicators and diagnostic criteria, and to examine their associations with mortality and life expectancy. METHODS: We analyzed data from 141,945 adults from a nationally representative cohort study in China. Cox proportional hazards regression models were used to estimate hazard ratios (HRs). Using fasting plasma glucose (FPG), 2-hour postload glucose (2hPG), and hemoglobin A1c (HbA1c) levels, prediabetes was defined according to the American Diabetes Association (ADA), World Health Organization (WHO), or International Expert Committee (IEC) criteria, and newly diagnosed diabetes according to the ADA criteria. RESULTS: Prediabetes prevalence varied widely across glycemic indicators, ranging from 3.0% to 26.1%, while newly diagnosed diabetes prevalence ranged from 2.6% to 4.4%. Over a median follow-up of 9.0 years, a total of 6,924 deaths were documented. Compared with people with normoglycemia, prediabetes defined by FPG (either ADA or WHO criteria) was not significantly associated with increased risks of all-cause or cardiovascular disease (CVD) mortality (all P>0.05). In contrast, prediabetes defined by 2hPG or HbA1c (either ADA or IEC criteria) was each associated with higher risks of all-cause mortality (HRs: 1.13-1.23; all P<0.001) and CVD mortality (HRs: 1.12-1.25; all P<0.001). Prediabetes defined by 2hPG or HbA1c, but not FPG, was associated with 1.1-2.3 years reduction in life expectancy, with the largest loss observed for IEC HbA1c. In addition, diabetes defined by FPG, 2hPG or HbA1c was each significantly associated with higher risk of all-cause and CVD mortality (HRs: 1.25-1.51), and reduction in life expectancy (2.0-3.7 years). Furthermore, the 2hPG-based definition of prediabetes and diabetes was associated with mortality risk, independent of FPG and HbA1c levels. CONCLUSIONS: These findings suggest that reliance on FPG alone may fail to identify certain individuals at elevated mortality risk. In contrast, 2hPG and HbA1c provide additional prognostic information beyond FPG.